Using Recent BCI Literature to Deepen our Understanding of Clinical Neurofeedback: A Short Review

Camille Jeunet, Fabien Lotte, Jean-Marie Batail, Pierre Philip, Jean‐Arthur Micoulaud FranchiNeuroscience 2018 · DOI

TL;DR

This paper bridges the gap between Brain-Computer Interface (BCI) and Neurofeedback (NF) research by suggesting that insights from BCI literature can significantly enhance our understanding and application of clinical NF. By adopting a human-factor-centered approach, the authors propose that adapting NF protocols based on task-specific and task-unspecific factors identified in BCI research can lead to more effective and individualized treatments.

What Problem Does This Paper Address?

The clinical efficacy of Neurofeedback (NF) is often debated, with significant variability observed between patients and studies. This variability is partly attributed to the complex interplay of factors influencing NF performance, which are not always well-understood or systematically addressed in clinical trials. While existing reviews have touched upon these factors, this paper aims to complement them by specifically drawing upon the rich literature in Brain-Computer Interfaces (BCIs) – a closely related field that shares common goals and underlying mechanisms with NF. The authors argue that by leveraging BCI research, particularly in areas like user training and adaptation, we can gain a deeper understanding of the processes underlying NF and improve the design of clinical trials, ultimately leading to more robust evidence for NF's efficacy.

Methods

This paper is a review and synthesis of existing literature, rather than a presentation of novel experimental methods. The authors adopt a "human-factor-centered approach," drawing parallels between BCI and NF training. They categorize influencing factors into three main groups:

  • Task-Specific Factors: These directly relate to helping the user self-regulate specific neurophysiological patterns. This includes adapting neurophysiological features (e.g., patient-specific frequency bands), feedback modalities, and mental strategies.
  • Task-Unspecific Factors: These indirectly influence the user's ability to self-regulate. They are further divided into:
  • Cognitive & Motivational Factors: Such as attention, motivation, and working memory.
  • Technology-Acceptance Factors: Including the sense of agency and user experience with the technology.

The authors propose that insights from BCI research, particularly in areas like machine learning for feature identification, adaptive feedback, and mental strategy optimization, can be directly applied to NF.

Datasets

Not discussed in the paper. This is a literature review, not an empirical study with datasets.

Key Insights and Ideas

  • BCI and NF are kindred spirits: Despite their differences, both BCI and NF aim to train users to modulate specific neurophysiological patterns, suggesting shared underlying cognitive and neurophysiological processes.
  • Human-factors are crucial: A human-factor-centered approach, considering the interaction between the user and the system, is essential for understanding and improving NF efficacy.
  • Task-unspecific factors matter: Beyond targeting specific brain rhythms, factors like motivation, attention, and the user's sense of control (agency) can significantly impact NF outcomes.
  • Machine Learning for Personalization: BCI research has extensively used machine learning to identify individualized features and optimize training, a practice that could greatly benefit NF by tailoring protocols to each patient.
  • Beyond Visual Feedback: Exploring alternative sensory modalities for feedback (e.g., tactile) and incorporating social/emotional elements (e.g., learning companions) can enhance user engagement and motivation in NF.
  • Mental Strategies are Adaptable: Just as in BCI, identifying and adapting optimal mental strategies for NF can improve self-regulation and performance.

Key Results

As this is a literature review, there are no quantitative experimental results. The "results" are the synthesized insights and proposed applications derived from the reviewed literature. The paper highlights that:

  • BCI machine learning tools can help design more individualized NF features (patient-specific frequency bands and brain areas).
  • Task-unspecific factors identified in BCI research (mu rhythm at rest, gamma power, sense of agency) may predict NF efficacy and help select appropriate NF protocols.
  • Tactile feedback, social presence, and positively biased feedback for novices (to enhance sense of agency) are promising avenues for NF, drawing from BCI successes.
  • Automatic difficulty adaptation and mindfulness meditation training, proven effective in BCI, could also benefit NF by promoting flow states and reducing anxiety.

Limitations and Future Work

  • Placebo Effect: The authors acknowledge the ongoing debate about the extent to which NF efficacy is due to a placebo effect. They suggest that while placebo effects may contribute, they should not oversimplify the issue, and that psychosocial variables are important.
  • Need for Rigorous Standards: The paper emphasizes the need for standardized NF training protocols and better assessment of learning processes in clinical trials.
  • Further Research: The authors call for more research to:
  • Identify the best mental strategies for NF and determine if specific strategies are universally useful.
  • Investigate the relevance of different sensory modalities for feedback depending on the NF paradigm.
  • Assess the impact of emotional support and social presence in NF protocols.
  • Explore how BCI-derived methods for adapting task difficulty and promoting flow states can be integrated into NF.
  • Investigate the role of cognitive/emotional factors and technology acceptance in NF training.

Summary updated

BibTeX
@inproceedings{jeunet2018using,
  author = {Camille Jeunet and Fabien Lotte and Jean-Marie Batail and Pierre Philip and Jean‐Arthur Micoulaud Franchi},
  title = {{Using Recent BCI Literature to Deepen our Understanding of Clinical Neurofeedback: A Short Review}},
  booktitle = {Neuroscience},
  year = {2018},
  doi = {10.1016/j.neuroscience.2018.03.013},
}
Figure